Structural and Catalytic Properties of S1 Nuclease from Aspergillus oryzae Responsible for Substrate Recognition, Cleavage, Non-Specificity, and Inhibition

Abstract

The single-strand-specific S1 nuclease from Aspergillus oryzae is an archetypal enzyme of the S1-P1 family of nucleases with a widespread use for biochemical analyses of nucleic acids. We present the first X-ray structure of this nuclease along with a thorough analysis of the reaction and inhibition mechanisms and of its properties responsible for identification and binding of ligands. Seven structures of S1 nuclease, six of which are complexes with products and inhibitors, and characterization of catalytic properties of a wild type and mutants reveal unknown attributes of the S1-P1 family. The active site can bind phosphate, nucleosides, and nucleotides in several distinguished ways. The nucleoside binding site accepts bases in two binding modes-shallow and deep. It can also undergo remodeling and so adapt to different ligands. The amino acid residue Asp65 is critical for activity while Asn154 secures interaction with the sugar moiety, and Lys68 is involved in interactions with the phosphate and sugar moieties of ligands. An additional nucleobase binding site was identified on the surface, which explains the absence of the Tyr site known from P1 nuclease. For the first time ternary complexes with ligands enable modeling of ssDNA binding in the active site cleft. Interpretation of the results in the context of the whole S1-P1 nuclease family significantly broadens our knowledge regarding ligand interaction modes and the strategies of adjustment of the enzyme surface and binding sites to achieve particular specificity.

Fig 1. Overview of the structure of S1 nuclease.

(a) S1 nuclease represented by its surface with important sites in sticks. The secondary structure representation is shown in the inset (helices are colored green, β–strands red, loops black). The catalytic zinc ions are shown as light blue spheres. Glycans are shown as sticks (carbon–green) and labelled according to the modified residues. Residues involved in the zinc cluster coordination are shown as sticks (carbon–green), residues forming NBS1, (also known as Phe–site) as sticks with carbon in yellow. Residues forming the Half–Tyr site are shown as sticks (carbon–magenta). NBS1 and the Half–Tyr site delimit the extent of the active site cleft. The graphics was created based on the structure 5FBF–nuclease products with removed 5’dCMP. (b) The active site of S1 nuclease. The color scheme is the same as in (a). Residues involved in the zinc cluster coordination and formation of NBS1 are labelled. The activated water molecule is labelled as W1, the water molecule present inside NBS1 as W_NBS1. Important interaction distances of W1, W_NBS1 and the distance of Asn154N^δ2 to Gly152O are given in Å. The graphics was created based on the structure 5FBA–phosphate, alternative without PO₄. Phe81 is shown in alternative A (occupancy factor 0.7). Molecular graphics were created using PyMOL (Schrödinger, LLC).

Fig 2. Comparison of amino–acid sequences and secondary structures of S1, P1, TBN1, and AtBFN2 nucleases.

Secondary–structure labeling for S1 nuclease is shown: α–helices as α, 3₁₀–helices as η, β–sheets as β, β–turns as T. Residues coordinating Zn²⁺ ions are marked by solid line, cysteine residues forming disulfide bridges by dotted line, S1 nuclease glycosylation sites conserved in P1 nuclease (Glyc. 2 is conserved in position in space, not in sequence) are marked by black arrows. Glycosylation sites not present in S1 are marked by empty arrows. Residues forming the Nucleoside binding site 1 (NBS1) are marked by diamonds. Labels of features conserved in all four nucleases are in bold letters. The Half–Tyr site of S1 nuclease is marked by stars. The Tyr site of P1 nuclease is marked by circles. Secondary structure was assigned by ENDscript [16] based on the structures of S1 5FBF–nuclease products, P1 nuclease–PDB ID: 1AK0 [11], AtBFN2 –PDB ID: 3W52 [13] and TBN1 –PDB ID: 3SNG [12]. The figure was created using ESPript [16] and manually edited.

Fig 3. Observed binding of the ligands in S1 nuclease structures.

The catalytic zinc ions are shown as light blue spheres. Asp65 (Asn65 in the case of the mutant) is shown in sticks (carbon–green). Other residues involved in the zinc cluster coordination are not shown. Lys68 and Tyr69 interact with ligands and are shown as sticks (carbon–green). Residues forming NBS1 are shown as sticks (carbon–yellow). Important water molecules are shown as red spheres and phosphate ions as orange/red sticks. Selected interactions are shown as black dashed lines. Molecular graphics were created using PyMOL (Schrödinger, LLC). PDB ID of each structure is shown. (a) Binding of water molecules in the unoccupied zinc cluster in the structure 5FB9 –unoccupied (pH 5.5). (b) The first binding mode of the phosphate ion in the structure 5FBB–inhibitors (pH 6.5). 5’AMP is excluded from the graphics for clarity. (c) The second binding mode of phosphate ion in the structure 5FBD–nucleotidase products (pH 4.2). All water molecules are displaced and none of the oxygen atoms of the phosphate ion occupies the original positions of water molecules. dCyt is excluded from the graphics for clarity. (d) The inverted deep binding mode of 5’AMP in the complex 5FBB–inhibitors (pH 6.5). Water W_NBS1 is replaced by the inhibitor. The zinc cluster is occupied by a phosphate ion. (e) The binding mode of 2'–deoxyadenosine 5'–thio–monophosphate in the complex 5FBC–remodeled (pH 5.5). The thiophosphate moiety binds outside the zinc cluster interacting only with Zn3 using its sulfur atom (marked as S), with Lys68 by the oxygen atom and through a water network with Asn154. In this deep binding mode W_NBS1 is replaced. Notice the same position of the adenine amino group, similar to the position described in (d), but an entirely inverted orientation of the base. NBS1 is remodeled to the extended form. (f) The observed binding mode of 2'–deoxycytidine (carbon–magenta) in the complex 5FBD–nucleotidase products. The cytosine moiety interacts with the protein using several types of interactions. Its π–conjugated system interacts with Phe81 and the peptide bond between Ala151 and Gly152. It has a direct polar interaction with the side chain of Asp83 and several water–mediated interactions, including involvement of water inside the NBS1 site. The 2’–deoxyribose moiety binds close to the zinc cluster and interacts directly with Lys68N^ζ, Tyr69O^η, Asn154N^δ2, and the phosphate ion. A similar binding occurs in the complex 5FBF–nuclease products (for the dCyt moiety of 5’dCMP) and in the complex 5FBG–mutant with products. (g) The observed binding mode of 2’–deoxyguanosine (carbon–magenta) in the complex 5FBG–mutant with products, chain B. Binding of the pyrimidine–like part of guanine mimics the orientation of cytosine inside NBS1 as shown in panel (f). N7 of the imidazole–like part is involved in the water network (W3 and W_dGua) connecting this atom to Zn3, the phosphate ion, and Lys68N^ζ. Asp154N^δ2 can interact with the π–conjugated electrons of the guanine moiety. (h) The observed binding mode of two molecules of 5’dCMP in the complex 5FBF–nuclease products (pH 4.2). The binding mode of the first 5’dCMP (carbon–pale pink) is almost identical with binding of 2’–deoxycytidine in the case of the complex 5FBD–nucleotidase products (panel f). The phosphate moiety is disordered and interacts either with Asn154 or with Lys68 and Tyr69. (i) The phosphate moiety of the second 5’dCMP in the complex 5FBF–nuclease products (carbon–magenta) binds in the zinc cluster in the second binding mode (as phosphate ion in 5FBD–nucleotidase products, panel f). The cytosine moiety likely interacts with Tyr69O^η (hydroxyl group) through its π–conjugated system.

Fig 4. Remodeling of NBS1.

The Nucleoside binding site 1 of S1 nuclease is represented as sticks (carbon–green). The solvent accessible surface of protein, color–coded by atom types, is shown. Zinc ions are shown as light blue spheres. Ligands are shown as sticks (carbon–grey). Selected interactions are shown as black dashed lines. The protein orientation in both panels is identical; changes of Phe81 and Asn154 can be seen. (a) The compact form of NBS1 with shallow base binding observed in the structure 5FBD–nucleotidase products. (b) The extended form of NBS1 with deep base binding observed in the structure 5FBC–remodeled. Molecular graphics were created using PyMOL (Schrödinger, LLC).

Fig 5. The Half–Tyr site of S1 nuclease and the proposed binding of ssDNA.

(a) 2'–deoxycytidine (shown as sticks, carbon–magenta) bound in the Half–Tyr site and comparison of the position of the Half–Tyr site of S1 nuclease (carbon–green) with the position of the Tyr site of P1 nuclease (carbon–dark blue) and with respect to the active site. Both secondary sites are positioned at the same end of the active site cleft, albeit on the opposite “banks”. Superposition of the complexes 5FBG–mutant with products, chain A, 5FBF–nuclease products, and of P1 nuclease (PDB ID: 1AK0 [11]) was calculated using the SSM Superpose tool in Coot [19]. Residues forming the Half–Tyr site are shown as sticks (carbon–green) and labelled. Residues of P1 nuclease involved in the formation of the Tyr site are shown as sticks (carbon–dark blue) and not labelled. The catalytic zinc ions (light blue spheres) with the phosphate ion (red/orange sticks) inside the zinc cluster are shown only for S1 nuclease. NBS1 is not shown but its position is marked. 2'–deoxycytidine and 5’dCMP binding in the active site cleft are shown (carbon–grey). (b) The proposed binding mode of ssDNA in the active site cleft based on the observed interactions of nucleotides and nucleosides in the S1 nuclease structures. The placement of the five nucleotides is based on x–ray structure coordinates except for the two nucleotides shown in green, which were positioned manually with optimized chain geometry in order to demonstrate the possible direction of ssDNA binding. Molecular graphics were created using PyMOL (Schrödinger, LLC).

Fig 6. Comparison of the catalytic activity of S1 wild type and mutants.

The catalytic activity of S1wt, S1D65N, S1K68N, S1N154A, and S1N154S is shown as a percentage of the S1wt activity on ssDNA.

Fig 7. Two types of interactions of the O3’ oxygen with the active site in the S1–P1 nuclease family.

Zinc ions are shown as light blue spheres. Asp65 (Asn65 in the case of the mutant) is shown in sticks (carbon–green). Other residues involved in the zinc cluster coordination are not shown. Lys68 and Tyr69 interact with ligands and are shown as sticks (carbon–green). Residues forming NBS1 are shown as sticks (carbon–yellow). Phosphate ion is shown as orange/red sticks, sulfate ion as yellow/red sticks, and selected interactions as black dashed lines. Molecular graphics were created using PyMOL (Schrödinger, LLC). (a) Comparison of the deoxyribose binding and position of its O3’ in the structure 5FBG–mutant with products with its position in the structure of AtBFN2 (PDB ID: 4CXO [14]). The zinc cluster, residues involved in the interactions and NBS1 are shown only for S1 nuclease. Ligands present in the structure of S1 nuclease are phosphate and dCyt (carbon–magenta). Ligands present in AtBFN2 are sulfate and thymidine 5’–monophosphate (carbon–silver). Notice the difference in the positions of the O3’ oxygens and of the phosphate ion in S1 –excluding the interaction of O3’ with Zn3. (b) Comparison of deoxyribose binding and the position of its O3’ in the structures 5FBF–nuclease products and AtBFN2 (PDB ID: 4CXO [14]). The zinc cluster, residues involved in the interactions and NBS1 are shown only for S1 nuclease. Ligands present in the structure of S1 nuclease are two molecules of 5’dCMP (carbon–magenta). Ligands of AtBFN2 are displayed as in panel (a). Notice the difference in the positions of the O3’ oxygens and of the phosphate moiety in the case of S1 which excludes binding of O3’ to Zn3.